Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Statistics›MCUSUM Chart — Multivariate CUSUM Control Chart
Process / pipelineStatistical process control

MCUSUM Chart — Multivariate CUSUM Control Chart

Multivariate CUSUM Control Chart · Also known as: Multivariate Cumulative Sum Chart, MCUSUM Control Chart, Crosier MCUSUM Scheme, Çok Değişkenli CUSUM Kontrol Grafiği

The Multivariate CUSUM (MCUSUM) Chart is a sequential monitoring scheme designed to detect small, persistent mean shifts in a process characterized by multiple correlated quality variables simultaneously. Introduced by Robert Crosier in 1988, it extends the classical univariate CUSUM principle to the multivariate setting by accumulating a vector-valued sum of deviations from the in-control mean, scaled by the process covariance structure, and comparing a scalar norm of that cumulative sum against a control limit.

ScholarGate
  1. Process / pipeline
  2. v1
  3. 1 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

MCUSUM Chart
CUSUM ChartMEWMA Chart

When to use it

Use the MCUSUM chart when monitoring several correlated process variables simultaneously and the primary concern is detecting small-to-moderate sustained mean shifts (roughly 0.5–2 standard deviations) rather than large abrupt changes. The method assumes the in-control mean vector and covariance matrix are known or accurately estimated from a Phase I study, and that observations are independent and follow a multivariate normal distribution. For very large shifts, the Hotelling T² chart may respond faster. When the shift direction is entirely unknown, MEWMA may offer comparable or superior power with simpler implementation.

Strengths & limitations

Strengths
  • Substantially faster detection of small, sustained mean shifts compared to the Hotelling T² chart due to its cumulative memory.
  • Explicitly accounts for correlations among quality variables through the covariance matrix, avoiding inflated false-alarm rates from separate univariate charts.
  • Intuitive shrinkage mechanism provides a clear geometric interpretation and prevents accumulation of noise-driven evidence.
  • ARL performance is well-characterized in the literature, enabling principled selection of design parameters k and h.
Limitations
  • Requires accurate knowledge of the in-control mean vector and covariance matrix; estimation error from small Phase I samples degrades ARL performance.
  • Assumes multivariate normality; robustness to heavy tails or skewness is limited, and violations can increase false-alarm rates.
  • Design parameter selection (k, h) typically relies on simulation or numerical tables rather than closed-form solutions, adding implementation complexity.
  • Does not directly identify which variable or combination of variables caused the signal, requiring post-signal diagnostic tools.

Frequently asked

How do I choose the reference value k and control limit h?

The reference value k is typically set to half the magnitude of the minimum mean shift (in standardized units) that the chart should detect quickly, a convention analogous to univariate CUSUM design. The control limit h is then determined by simulation or published ARL tables to achieve the desired in-control ARL, commonly 200 or 370. Crosier's 1988 paper provides initial ARL results, and subsequent simulation studies offer broader tabulations.

How does MCUSUM compare to MEWMA for detecting small shifts?

Both charts offer substantial improvement over Hotelling T² for small shifts. MEWMA (multivariate EWMA) is often slightly easier to implement and tune, while MCUSUM's shrinkage mechanism provides a clear geometric interpretation. ARL comparisons in the literature show the two are broadly competitive, with performance depending on the shift direction, magnitude, and the chosen design parameters.

Can MCUSUM handle non-normal data?

The standard MCUSUM chart is derived under multivariate normality, and its ARL properties depend on this assumption. For moderately non-normal data, the chart can still function but with degraded ARL guarantees. Researchers have proposed nonparametric and rank-based MCUSUM variants for settings where normality cannot be justified; these should be considered when distributional assumptions are questionable.

Sources

  1. Crosier, R. B. (1988). Multivariate generalizations of cumulative sum quality-control schemes. Technometrics, 30(3), 291–303. DOI: 10.2307/1270083 ↗

How to cite this page

ScholarGate. (2026, June 2). Multivariate CUSUM Control Chart. ScholarGate. https://scholargate.app/en/statistics/mcusum-chart

Related methods

CUSUM ChartMEWMA Chart

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

  • CUSUM ChartStatistics↔ compare
  • MEWMA ChartStatistics↔ compare
Compare side by side →

Referenced by

MEWMA Chart

Similar methods

MEWMA ChartMulti-response Control ChartCUSUM ChartMulti-response statistical process controlHybrid Control ChartEWMA ChartHybrid Statistical Process ControlSensitivity Analysis with Control Chart

Related reference concepts

Multivariate Analysis of VarianceStatistical Process Control and Run ChartsPrincipal Component AnalysisMultivariate RegressionMultivariate Multiple RegressionMultivariate Analysis

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — MCUSUM Chart (Multivariate CUSUM Control Chart). Retrieved 2026-07-21 from https://scholargate.app/en/statistics/mcusum-chart · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Robert Crosier
Year
1988
Type
Multivariate sequential monitoring chart
Subfamily
Statistical process control
Signal Statistic
Cumulative sum of standardized multivariate deviations
Reference Distribution
Multivariate Normal
Related methods
CUSUM ChartMEWMA Chart
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account